LATTICE: A Framework for Optimizing IoT System Configurations at the Edge

Abstract

The Internet of Things is expected to contribute to a "smarter world" by connecting the physical to the virtual, i.e., enabling advanced knowledge engineering over the big data gathered about the physical world. However, such a promise comes along with high resource consumption, spanning the network, storage and computational resources, not to mention possible security and privacy threats. As a result, it tends to be admitted that the IoT smartness will not be accommodated at scale by a centralized cloud-based approach. Instead, the deployment of IoT systems needs to leverage a highly distributed system architecture, which optimizes the distribution of the computation-from the edge to the cloud-according to the unique business requirements in terms of financial cost, latency, availability, etc. Toward that goal, this paper introduces the LATTICE framework, which aims at taming the complexity of configuring edge-based IoT systems. LATTICE builds upon ontologies that have proven useful to characterize the constituents of IoT systems in the required domain-specific way. However, LATTICE also revisits the exploitation of ontologies-i.e., the formal description of the real world, spanning the physical and cyber entities-across the development life-cycle of the IoT systems. As a first evidence, this paper introduces an automated approach to the optimization of IoT system configurations at the edge, provided the ontological description of the target IoT system.

Adaptive Scheduling of Continuous Operators for IoT Edge Analytics

“…The terms Edge and Fog computing are often used interchangeably [32]. However, in this work we consider them as different layers where the boundary between the two is tiny [33].…”

Design and Implementation of a Hydroponic Strawberry Monitoring and Harvesting Timing Information Supporting System Based on Nano AI-Cloud and IoT-Edge

“…During automation, the levels of pH, water, temperature and humidity are adjusted. Issarny et al [11] introduced the LATTICE framework for the optimization of IoT system configuration at the edge, provided the ontological description of the target IoT system. The framework showed an example applied to a hydroponic room of vegetables for monitoring and controlling several physical variables (e.g., temperature, humidity, CO 2 , air flow, lighting and fertilizer concentration, balance and pH).…”

“…The proposed system collects 13 types of strawberry growth environment data, whereas the environmental information monitored in the previous related works [5][6][7][8][9][10][11][12][13] consisted of two to eight types. By collecting more environmental information compared to the related works, it is possible to access more diverse methods when analyzing the growing environment.…”

“…By collecting more environmental information compared to the related works, it is possible to access more diverse methods when analyzing the growing environment. As the related works [5][6][7][8][9][10][11][12][13] involve simply collecting environmental data or simply storing the collected data in the cloud, the addition of functions to the related works is limited. By contrast, the proposed method is designed to facilitate function expansion and analysis, as it consists of an IoT-Edge module and a nano-sized private AI-Cloud module.…”